Philosophers, historians, and social scientists have proposed a multitude of different theories trying to explain the rise of huge complex human societies over the past few millennia. Was the primary driver the invention of agriculture, which seems to be the default explanation held by many archaeologists? Or was it private property and class oppression, as many Marxists believe? Warfare between tribes was a popular explanation a century ago, and has been recently revived under the rubric of cultural multilevel selection. Was it large-distance trade, the need for sophisticated information management, or something else?
One of the main goals of the Seshat project is to answer these sort of Big Questions. Nor are these questions only of interest to scholars who like to argue back-and-forth over esoteric details. Issues like the nature and causes of human social development are as relevant to people today as they were for the historical societies we study. Answering these Big Questions points the way to understanding how we arrived at our current state of affairs, and history may also hold the key to avoiding the sort of mistakes that doomed many previous societies.
We went about exploring this critical issue by gathering together massive amounts of data about past societies. We have just taken an important step towards this goal with the publication of the first article that fully utilizes the power of the Seshat approach: “Quantitative historical analysis uncovers a single dimension of complexity that structures global variation in human social organization,” published recently in the Proceedings of the National Academy of Sciences (PNAS).
In this paper, we analyze data describing over 400 past “polities” (ranging from large ones, like states and empires, to small chiefdoms, and even politically independent villages). Our data comes from all world regions (see map here) and extends back in time to the origins of agriculture, in many cases going back thousands of years into the past.
Frequency distribution of the starting dates for data sequences in Seshat Databank. For 10 global sampling points (“late complexity”, see map) data series are short, often starting only when European explorers reached the area in the eighteenth or nineteenth century. For “early complexity” locations data sequences extend back in time between 4,000 and 10,000 years ago. “Intermediate complexity” cases are between these two extremes.
These data provide a truly unprecedented scale of historical material to work with, making our statistical analyzes quite powerful.
The PNAS article does not yet directly answer the questions we posed in the beginning of this blog post. Rather, it’s a foundational article. In order for us to test theories about what caused the evolution of social complexity, we first need to answer a very basic question: what is social complexity? We need to first know what we want to explain before we can go about tracing its evolution.
What we were able to accomplish is to identify nine distinct key characteristics that together measure the developmental trajectories of all the societies used in our analysis (see figure below).
Nearly all previous attempts to explain social complexity focus on only one or two aspects, like the size of the largest settlement, or the number of levels in administrative hierarchies. Our unique analysis, tracking a much larger number of variables within a huge sample of historic polities, was able to discover that, in fact, social complexity involves the co-evolution of these nine characteristics together. In taking the first steps to quantify the development of social complexity, the PNAS article provides a very solid, scientifically sound way for us to now answer the more interesting questions: what causes societies to gain or lose this complexity, and what are the consequences of changes in complexity on other aspects of societies?
We have already started putting this result to work. Several other articles are already going through their paces in academic journals (but the millstones of scientific publishing grind very slowly) or are about to be submitted. We have also published the data on which the PNAS paper is based, so that others can participate in the fun.
You can download the dataset suitable for statistical analyses here. Currently the dataset is available as a downloadable spreadsheet. In the coming weeks, we will be uploading our entire Social Complexity dataset onto our project website in a more easy-to-use, viewable format; keep checking the site for updates!
Peter Turchin and Dan Hoyer
This is a repost of The first article to utilize the full power of the Seshat: Global History Databank has arrived!
What do you think about Alexandre Deulofeu’s works and his “Mathematics of History”?
I suppose you know the work of this author and his 17-century historical cycle. Does it make sense to you from a client dynamic point of view?
Deulofeu also provides a process of empire formation with two imperial moments, a first of federal character followed by a second of unitarist (absolutist) character. Deulofeu envisages a duration of 550 years for this imperial cycle. This allowed him to predict the fall of the USSR 30 years in advance. What do you think about it?
Following this reasoning Deulofeu announced that around 2029 the Spanish empire would have disappeared. This disappearance means that the Spanish empire would have returned to a situation similar to the one prior to the first imperial moment. In other words, the Iberian peninsula would be divided into different territories that are politically independent from each other, similar to the situation that existed prior to the union of Castile and Aragon.
According to Deulofeu, the main cause of the decadence of empires is the corruption of the elites more interested in accumulating wealth in any form (legal or criminal) than in governing their citizens equitably and seeking economic development that benefits everyone.
Curiously (or not) in 2010, the government of Madrid broke the constitutional pact of 1978 and judicially imposed a statute of autonomy on the Catalans who did not vote. The reaction, as we know, has led to a notable increase in independence in Catalonia, which has gone from a testimonial 15% to the current 47.5%. The central government from Madrid in the last 7 years has been making all kinds of mistakes that have only fueled this trend.
The majority of Catalan citizens are convinced that the central government of Madrid is the most corrupt in the history of Spain and possibly Europe. At the moment the government party has more than 900 accused of corruption, all the members of the ruling party since it has existed are or have been charged, the same president of the government M. Rajoy appears in the documents of the last treasurer charged by the party as beneficiary of undeclared bonuses collected in envelope to avoid the payment of taxes. It seems, then, that the “Spanish empire” is irreformable precisely because if the entire current corrupt system were to be reformed, it would fall like a house of cards, putting a large part of the current “stables”in prison.
How could Deulofeu accurately predict this type of event all by 1967?
I am not as familiar with Deulofeu’s work as I wish, but his explanation of imperial decline that focuses on the elites is very similar to the elite overproduction principle in the structural-demographic theory
http://peterturchin.com/age-of-discord/
“ll the members of the ruling party since it has existed are or have been charged”
Sorry, I wanted to say, until today, all accountants of the ruling party have been charged with various crimes.
It’s a niggle about a very exciting preview (will read), but does any evolutionary process have “a cause” rather than “conditions [of existence] within which” it happened?
Causality is of course a complex subject, but if you are fine with in physics, it equally applies to evolution. In particular, processes exerting selective pressure on traits under evolution are surely causal, no?
So in a famous example of industrial melanism, the causal factor in the evolution of dark coloration in moths is the great amounts of soot on tree trunks where moths sit during the day.
Certainly there is a primary cause/ but it also creates conditions, which on their own influence the distribution of the cause.
In general existence depends on a time/energy equation, which creates a neutralization girdle for it, where demand (supposed shortage, which can also be knowledge) meets the opposite energy and reaches singularity. Primarily this was the time to get food (energy) before it is missing.
Which is very well possible because our own biological (8h labor) time is much shorter than the one of the earth (24h) and that of the seasons.
Hunter gatherer societies took all their time all year long gathering food/ while agricultural only lasts untill harvest season. They have gained a lot of free time for next needs. The lead for a very different evolution of human kind.
Another consequence of it was the creation of money; because the food was owned long before it was used as energy in time. Money tokens represented food/ but could be used for trade independent from food. It started to represent a supposed value. Which was later on swapped for gold/ but with its own new next consequences, because gold does not grow like food, when an economy needs to grow. Hence the bancruptcy of bankers in economic crisis, because they printed too much money relative to the real economy, especially when it became more efficient and needed less time/energy. The inverted consequence plays out in its money system, because it cannot adapt. In evolution theory a well known disruption of possibilities for a kind to survive. A condition still haunting us in our current speculative economies.
So you see it depends on believe systems. Which humanity created himself. ‘As biology creating its own environmental conditions, as the cause for new evolution.’
Societies of mankind basicly are economic time equations with their own enclosed specializations. Professions in stead of biological kinds. Therefore they as a group potentially have a collective inteligence of some.kind. This can also be explained as altruism/ although it is the swap of individual interests in general, which represent the same on another time scale.
For instance at the same time the general interest of society as a whole, is a broader one than its limited material economy. This is represented in the federal state or democracy, thinking up general rules: hierarchy/ but also depending on its own understanding of functioning and its economic money income. Which is peculiar, since it postpones new labor, depending on the proceeds of other labor, which it does not depend on anymore. Hence complicated insurances. Apparently humanity still has not understood his own economy.
It creates the entanglement of politicians with neo libertarians, promising a thriving future/ but mostly in their own interest = peculiar understanding of a believe system.
Opposing elites in the past were often either monarchist military interests/ or those involved in trade (fi Spain, Netherlands in history). It depended on the sum of the potential, which one was more promesing or costly.
Also the ownership of land resulted in capital, able to buy labor/ but in fact only bought consumption before labor. Hence the ‘believe system’ created by its end condition, that of wage slavory.
In that sence Marxists were correct in their observation/ but it still depends on the design of money, economy and its evolutionary stage, whether an alternative concept is succesfull or not.
It was a very pleasant surprise to discover that a Seshat-based paper was finally published. It was even more pleasant to discover that it is open access. I could read it and look at its methods. Like filling in missing data by extrapolating from known data, and doing simulations with multiple random fill-ins.
That there was only one strong principal component is a very interesting result — and a robust result, not much affected by different ways of handling missing data. But it’s a correlation, and it does not indicate anything about the directions of causality.
It would have been interesting to find more strong dimensions of variation, but all the others are very weak.
Hi Loren — thanks. And I share your frustration with the slow grinding of academic journals. But that’s the game in town.
We are already taking next steps, and there are several articles in the works. In particular, I have just finished writing a methodological paper addressing just the issue you bring up: how do we use Seshat data to test causality hypotheses. This paper will be available soon, one way or another (e.g., by posting its preprint on SocArxiv).
Looking at the paper’s supplementary information, I found the data’s second principal component. It is not much stronger than the others; it does not stand out the way that the first one does.
It has an opposition of size vs. information. Size: polity population, polity territory, capital population, and levels of hierarchy, with loading of -0.21 to -0.30. Information: writing, text, money, with loading of +0.23 to +0.34. The rest: government services and infrastructure, with loading of +0.08 to +0.09, thus being a little on the information side. So as societies become larger and more complicated, they can grow more in one or the other, though that is not a strong effect.